estimate.autocorr.m: Calculate sampling interval based on exponential...
In danlwarren/RWTY: R We There Yet? Visualizing MCMC Convergence in Phylogenetics

Description Usage Arguments Value Examples

This function uses an exponential semivariogram model to estimate the asymptotic topological distance, and uses that to estimate the sampling interval at which topological distances have reached some fixed proportion of that value (default 0.95). It expects as input a data table output by rwty's topological.autocorr function

1	estimate.autocorr.m(dat, ac.cutoff = 0.95)

`dat`	A data frame output from topological.autocorr.
`ac.cutoff`	Default 0.95. The proportion of the asymptotic topological distance to use as a cutoff for determining sampling interval. For example, if ac.cutoff = 0.9, then the minimum sampling interval returned is the one that guarantees a topological distance at least 0.9 times the asymptotic value.

A data frame consisting of the value matching the ac.cutoff proportion of the asymptotic topological distance for each chain. This sampling interval estimates the interval at which topological distances are no longer autocorrelated. If the value is larger than the largest sampling distance, the table records this as a value of -1

data(fungus)
## Not run: 
# To get a good estimate we need all sampling intervals
autocorr.intervals = as.integer(length(fungus[[1]]$trees)/21)
sampling.table <- topological.autocorr(fungus, burnin = 20, autocorr.intervals = autocorr.intervals)
estimate.autocorr.m(sampling.table)

## End(Not run)